<p>This is not directly a DNS problem, it’s a network routing problem between some parts of the internet and the DNS servers for <a href="http://serverfault.com">serverfault.com</a>. Since the nameservers can’t be reached the domain stops resolving.</p>
<p>As far as I can tell the routing problem is on the (Global Crossing?) router with IP address <code>204.245.39.50</code>.</p>
<p>As <a href="https://serverfault.com/questions/42678/dns-failing-to-propagate-worldwide/42686#42686">shown</a> by <a href="https://serverfault.com/users/9503/radius">@radius</a>, packets to ns52 (as used by <a href="http://stackoverflow.com">stackoverflow.com</a>) pass from here to <code>208.109.115.121</code> and from there work correctly. However packets to ns22 go instead to <code>208.109.115.201</code>.</p>
<p>Since those two addresses are both in the same <code>/24</code> and the corresponding BGP announcement is also for a <code>/24</code> this <em>shouldn’t happen</em>.</p>
<p>I’ve done traceroutes via my network which ultimately uses MFN <a href="http://Above.net">Above.net</a> instead of Global Crossing to get to GoDaddy and there’s no sign of any routing trickery below the <code>/24</code> level - both name servers have identical traceroutes from here.</p>
<p>The only times I’ve ever seen something like this it was broken <a href="http://en.wikipedia.org/wiki/Cisco_Express_Forwarding">Cisco Express Forwarding</a> (CEF). This is a hardware level cache used to accelerate packet routing. Unfortunately just occasionally it gets out of sync with the real routing table, and tries to forward packets via the wrong interface. CEF entries can go down to the <code>/32</code> level even if the underlying routing table entry is for a <code>/24</code>. It’s tricky to find these sorts of problems, but once identified they’re normally easy to fix.</p>
<p>I’ve e-mailed GC and also tried to speak to them, but they won’t create a ticket for non-customers. If any of you <strong>are</strong> a customer of GC, please try and report this…</p>
<p><strong>UPDATE at 10:38 UTC</strong> As Jeff has noted the problem has now cleared. Traceroutes to both servers mentioned above now go via the <code>208.109.115.121</code> next hop.</p>